Psychometric testing of two measures of vaccination attitudes among U.S. Latine respondents

Although attitudes toward vaccination predict vaccination engagement, widely used measures of vaccination attitudes are untested and unvalidated among diverse populations. Latine communities experience disproportionately poor COVID-19 and other communicable disease outcomes. The objective of this study is to test psychometric properties of widely used vaccine hesitancy and vaccine acceptance measures among diverse U.S. Latine populations.

We conducted a psychometric meta-analysis of 13,406 Latine survey participants between 2021 and 2023 from 10 U.S. Rapid Diagnostics for Underserved Populations Initiative sites. Separate vaccine hesitancy and vaccine acceptance checklists were evaluated using two-parameter logistic item response theory modeling to assess item difficulty, discrimination, and measurement invariance across 6 pandemic waves. External validity was tested with vaccination status and vaccine course completion.

Vaccine acceptance items performed well in terms of measuring persons low to high in the trait and performed well at differentiating respondents. Generally, the vaccine hesitancy items displayed greater range and variance across the trait of hesitancy (item difficulties) than vaccine acceptance items. Equal slope discrimination could be assumed for the hesitancy items across pandemic waves, but scalar invariance was rejected, indicating that level of endorsement (item difficulties) increased over waves. Both metric invariance and scalar invariance were rejected for vaccine acceptance items, indicating both discrimination and difficulty increased across waves. External validity was supported for vaccine acceptance but not hesitancy. The point biserial correlation for vaccine acceptance was significantly associated with vaccination status (r = 0.53, p < .001) and course completion (r = 0.09, p < .001).

Item characteristics suggest both vaccine acceptance items and hesitancy items exhibited acceptable measurement properties, but vaccine acceptance met thresholds for consistency and external validity. This is the first psychometric testing of these vaccination attitudes measures, adding to validated measures for future use in Latine populations. Moreover, findings corroborate growing evidence of the dissociation between acceptance and hesitancy as constructs related to vaccination attitudes.